Problem: What is the most effective way to tell if a person is lying?
Hypothesis: If a person is lying, you will most accurately be able to tell by a spike in their blood pressure.
Step-by-step procedure
Recrute 10 participants ages 13-19 and collect their signed consent forms prior to beginning any part of the experiment.
Record the standing blood pressure, notes on facial features and body language, and notes on speech of your participants before beginning the questionaire, this will be the control.
Then, provide each of them with the first sheet of questions. After they have read over them, instruct them to lie for the answer on three of the questions.
From here, individually question each participant individually. This will happen in three rounds. During the first round of questioning, monitor the particiants blood pressure as they answer each question.
From the information you are collecting, decide whether or not you believe the person is lying or telling the truth, Document this.
At the end of the first round of questioning, have the participants reveal which answers they are lying about. Compare these answers with your initial determination.
Repeat these steps for rounds two (which will focus on facial features and body language) and three (which will focus on speech). Each round will require a new set of questions.
The success rate of each technique should be calculated at the very end of all questioning to determine which is most successful. To do this, for each technique take the number of lies you correctly spotted and divide it by the total number of lies for the round (which will be 30). You will recieve a decimal, multiply this by 100 and round to the nearest hundreth to come to your concluding percentage.
Once each percentage of success is calculated, determine the highest success rate. This is the most effective technique.
Materials List (specific)
10 participants ages 13-19
an informed consent form for each participant
ReliOn blood pressure/ pulse monitor
a knowledge base or reference on facial movements and features associated with lying
3 questionaires- each 10 questions long. A copy of each for each participant.
A calculator to be able to determine success rate.
Abstract:In this experiment, I attempted to answer the question of "what is the most effective way to detect lying?" To do so, I needed to determine several common lie detectors. To test them, 3 different techniques of lie detections are compared. This is done through a series of questions being asked to 10 participants in 3 sets- one for each technique. By using each technique, I determined which answers were lies. After the actual lies were revealed, the success rate of each technique was determined and they were all compared to determine the most effective technique.
Log Book 11/17/11 -
Went to Staples to purchase printer paper
Printed 11 consent forms
Intended for distribution tomorrow and return on monday
11/18/11
Acquired the blood pressure monitor needed for trials.
11/22/11
Distributed consent forms on Friday and Monday, now have collected them today.
Set up a date for as many trials as possible with as many participants as possible.
Date being Sunday, 12/11/11. Available participants that day:
Bryan Haynie
Ellie Lyons
Kirsten Lyons (possibly)
Graeme Nelson (possibly)
Also, i realized that I do not have tables for control statements, so I have placed them below with my questioning tables.
12/11/12:
Participants tested this day include:
-Marley Vebares
-Maddie Arnold-Scerbo
-Aoife Iaria
Data compiled on notebook paper to later be organized in the data table.
Today, I also decided that comparing the shifting blood pressures was very difficult. I am not 100% sure of how to tell a shift from a decline with respect to the two numbers. For these reasons, I changed the "Blood Pressure" sections to "Pulse".
Data compiled on notebook paper to later be organized in the data tables.
12/20/11:
Confirmed testing date of 12/23/11 for:
-Erin Gildea
-Mason Zeplo
-Klaudia Edwards
They are confirmed to be coming to my house after the half day of school on Friday the 23rd, for testing purposes.
12/21/11:
Today during class, I was hit with the realization that I now have to change my hypothesis due to the changing of my data types. I changed my "Blood Pressure" section to testing a person's pulse. so now my new hypothesis is: If someone is lying, then it will be most accurately detected by a spike in their pulse.
12/23/11:
Today, I tested the following participants:
-Erin Gildea:
-Klaudia Edwards
-Mason Zeplo
-Graeme Nelson
I will be sorting their results into the graphs at a later date.
1/2/12:
Today, I tested my final participant Kirsten Lyons. Her information will be placed into the graphs at a later date.
1/5/12:
Today, I will be uploading as much data into the charts as possible during the class period. Everything else will be finished after school today and this weekend. Unfortunately, I believe that I still have too much to finish to be able to turn in my finished product tomorrow for the extra credit, but the weekend will give me plenty of time to finish it up and turn it in on Monday at the beginning of class.
During the period today, I accidentally put the information in the wrong graphs, so I will have to fix it at home today.
1/5/12: Fixing the incorrect data tables, calculating percents, doing some organization and beginning to write all of my discussions and conclusions. Also, I calculated my success rates and compared them. Facial Idiosyncracies proved to be the best.
1/7/12: Today, I will be finishinng my project! I will be typing all of my final conclusions, dicussions, and analysises, orgainizing, and uploading my figures (graphs and pictures).
I am dealing with the formatting the best I can at this point. Everytime I save, it goes back to the way it does before I last saved.
Data TablesT= TrueL= Lie Table 1: Control Statements from all 10 participants
Participant
Standing Pulse (BPM)
Facial/ bodily idiosyncrasies
Comments on Speech
Ellie Lyons
87
*Avoids eye contact *Touches her face
*Hesitates to speak. *Makes noises to hide laughter
Kirsten Lyons
86
*Touches face *Looks around the room intently
*tends to make noises to shield discomfort
Erin Gildea
85
*Looks intently around the room *Raises eyebrows
*speaks quickly in a high voice
Graeme Nelson
79
*Covers mouth with hands *Leans forward
none
Klaudia Edwards
83
*twirls hair often
*hesitates (lots of 'well...." and "um...")
Marley Vebares
82
*Squints when lying
*Giggles during lying *Mumbles/ slurs words
Maddie Arnold-Scerbo
80
None
*Giggles after an attempted lie
Mason Zeplo
83
*Avoids eye-contact *fidgets with hands
*Speaks loudly when he is lying
Aoife Iaria
80
*Smiles when lying *avoids eye contact
None
Bryan Haynie
89
*I assume the high pulse is because the participant had consumed a Mountain Dew soda within an hour of testing. *Smiles and laughs when lying.
The most successful test was Facical Idiosyncracies, which proved my hypothesis wrong.
Analysis
Experimental Background:
To lie is to present a statement that is not the complete and whole truth. The average person lies several times a day about a wide variety of things. The spectrum of lying ranges from little white lies to deep, malicious lies intended to hurt someone. Testing a lie can be done in several ways. These may include: most popularly, a polygraph test, analysis of body language or facial idiosyncracies, or the way the speak and explain themselves. None of these are a definite way to spot lying. A professional grade polygraph test has a 40% margin of error. Neither of the other trials even have a set rate of error or success because they are so difficult to test due to the fact that they are situational.
My project tested the accuracies of several lie detection techniques. I tested these less-used techniques against a make-shift polygraph test (blood pressure/ pulse monitor). I rounded up 10 test subjects between the ages of 13-19 years old and tested their lies against 3 variables: Facial idiosyncracies and body language, speech, and pulse. Each variable was tested with a set of 10 questions. The test subjects were prompted to lie for 3 questions per set of 10. I then had to use each of these techniques to determine which answers were lies upon my questioning them. According to my research, the pulse test should have been the most accurate, while in fact, it was the least.
Discussion: This experiment was designed to compare three lie detectors and determine which was most successful. By testing the techniques of facial/body language, speech, and pulse, I determined that facial and body language are the most successful way to determine lying. You can see this at the end summarizations for Tables 3, 5, and 7 and Figure 1. This, although, goes against all of my research which stated that pulse and blood pressure would be most accurate. In fact, the pulse test proved to be the least effective by far, which can be seen in the summarization of Table 7 and Figure 1. My reasoning for being most accurate for facial and body language are that I knew my participants very well. I am accustomed to their idiosyncracies and their lying patterns. My results may have been drastically different if I had been testing people that I have never met. At that point, I would have to completely rely on my very basic knowledge of lying and what it does to face and body language. Instead, I could use prior knowledge of my participants to aid me in my decision making process. Because of this, my data may have been a little clouded. My data may have also been roughed up because of the quality of my equipment. My pulse monitor way by no means professional-grade (Figure 2). It often took a few tries to get a reading on a participant's pulse. Because of this, it was also easily manipulated. After my participants found this to be true, they silently began taking advantage of it. As you can see when comparing Tables 6 and 7, I sometimes did not detect any lies from participants. It was very hit or miss. If I were to redo my experiment, I would definitely invest in a more proper pulse monitor. Another possible margin of error could be that I let my test subjects review and hold the question sheets before and while I was questioning them (Figure 3). This way, they had time to decide which they were going to lie for and how. They did not have to lie on the spot, which made it easier to manipulate their features, body language, and breathing (which effected their pulse). This may have effected my data in more of a way than I had originally predicted. I would also change this about my project to improve it. The added stress of having to come up with a lie on the spot may have an affect on the subject's pulse, features, and body language. Depending on how much of an impact this had on each person, it could change my end results completely. The final thing I would change to improve the project is that I would test more participants or expand my age group. Obviously, by testing more people, I could achieve a more accurate average for each technique. But for the purposes of this project, I believed 10 would be enough. If I were to continue this project next year, I would like to see if testing a bigger demograpphic would vary my end results (which can be seen in Figure 1) Originally, I chose the age group that I did for reasons that applied to the way I was originally going to go about the project. But as I changed my ideas it became irrelavent to my question. Maybe it is more difficult to trigger a spike in blood pressure in young people as opposed to an older age group. They were all relatively close together (Table 1) and did not vary much through my testing. That could be another aspect reasoning for not being able to detect a majority of the lies in the pulse section and its surprisingly low success rate of 40% (Seen in Tables 6 and 7 and Figure 1). I predicted that speech and face/ body language would be close to each other. But as I documented my control statements (Table 1) and realized that few of my participants had definite speech ticks. It was here that I began to think that speech would fall just behind facial ticks, which it did. But, if I did not know my participants so well, they could possibly have lied about their control statements. That could be a potential margin of error for testing stangers, and I would definitely make that a part of my consent/ agreement form if I were to take that step with this experiment later. To conclude, by testing the techniques of facial ticks and body language, speech ticks, and pulse spikes on 10 participants, I found that the best way to detect lying is through facial ticks and body language (as seen in Figure 1). There were several flaws in my experiment that I would change if I were to continue, and I would also invest in more professional equipment. Now that I already have a base in the experiment, my predictions for its continuation stand with my final results here. These can be seen in Figure 1 and the math is shown in the summarization after Table 7. Application: The results in the experiment could be very beneficial. That ranges from a Detective or police officer, a lawyer, to someone who is just looking to spot a lie in a story. By knowing how to effectively tell if a person is lying to you, it is easiest to be able to find the ticks that confirm or deny their statement. Even more, if you personally are trying to lie your way out of a situation, you would know what to avoid as to not be caught, such as someone in the who has gone undercover. The study could go further to other ways of detection which could be compared to the ones tested here. That way, there would be even more of a variety of accuracy. All in all, this experiment can pertain to anyone who ever has needed to spot a lie or has ever needed to lie their way out of a situation. Conclusion: My hypothesis stated that the most effective way to spot a lie was that their pulse would spike. As you can see through all of my data tables and Figure 1, this hypothesis is incorrect. The most accurate way to spot a lie, as determined by this experiment, it a person's facial ticks and body language. After testing 10 participants on all three of the techniques, facial ticks had the highest success rate by far. The percent of success being 83.33%, 66.67% for speech, and 40% for pulse. Despite the flaws in this experiment, it is safe to conclude that by analyzing a person's facial features and body language, it will most effectively detect if a person is lying or not. Figures Figure 1
I'm not sure why it wont let me just have the picture of the graph in here but I suppose that it is because this is a Pdf file. It opens to the graph when you double click on it. It is my success rate comparison graph. Figure 2
The blood pressure and pulse monitor used for my experiment
Figure 3
Test subject Ellie Lyons going over the questions for the upcoming section 2: Speech Sources:
Here is my list of sources. It is having a lot of formatting issues and is not saving the way that I am trying to organize the information, so I typed it in googledocs and uploaded the Pdf of it. It wil open just like the graph.
Ps: This is not the only place where it refuses to save the formatting. It changes font size, color, and placement all over the document.
Title: Lying
Problem: What is the most effective way to tell if a person is lying?
Hypothesis: If a person is lying, you will most accurately be able to tell by a spike in their blood pressure.
Step-by-step procedure
Materials List (specific)
Abstract:In this experiment, I attempted to answer the question of "what is the most effective way to detect lying?" To do so, I needed to determine several common lie detectors. To test them, 3 different techniques of lie detections are compared. This is done through a series of questions being asked to 10 participants in 3 sets- one for each technique. By using each technique, I determined which answers were lies. After the actual lies were revealed, the success rate of each technique was determined and they were all compared to determine the most effective technique.
Log Book
11/17/11 -
- Went to Staples to purchase printer paper
- Printed 11 consent forms
- Intended for distribution tomorrow and return on monday
11/18/1111/22/11
- Distributed consent forms on Friday and Monday, now have collected them today.
My participants:-Ellie Lyons- 16
-Kirsten Lyons- 17
-Erin Gildea- 15
-Klaudia Edwards-15
-Marley Vebares-15
-Maddie Arnold-Scerbo-16
-Mason Zeplo-15
-Graeme Nelson-16
-Aoife Iaria-15
- Bryan Haynie-13
12/5/11
- Set up a date for as many trials as possible with as many participants as possible.
Date being Sunday, 12/11/11. Available participants that day:Also, i realized that I do not have tables for control statements, so I have placed them below with my questioning tables.
12/11/12:
Participants tested this day include:
-Marley Vebares
-Maddie Arnold-Scerbo
-Aoife Iaria
Data compiled on notebook paper to later be organized in the data table.
Today, I also decided that comparing the shifting blood pressures was very difficult. I am not 100% sure of how to tell a shift from a decline with respect to the two numbers. For these reasons, I changed the "Blood Pressure" sections to "Pulse".
12/18/12:
Participants tested:
-Ellie Lyons
-Bryan Haynie
Data compiled on notebook paper to later be organized in the data tables.
12/20/11:
Confirmed testing date of 12/23/11 for:
-Erin Gildea
-Mason Zeplo
-Klaudia Edwards
They are confirmed to be coming to my house after the half day of school on Friday the 23rd, for testing purposes.
12/21/11:
Today during class, I was hit with the realization that I now have to change my hypothesis due to the changing of my data types. I changed my "Blood Pressure" section to testing a person's pulse. so now my new hypothesis is: If someone is lying, then it will be most accurately detected by a spike in their pulse.
12/23/11:
Today, I tested the following participants:
-Erin Gildea:
-Klaudia Edwards
-Mason Zeplo
-Graeme Nelson
I will be sorting their results into the graphs at a later date.
1/2/12:
Today, I tested my final participant Kirsten Lyons. Her information will be placed into the graphs at a later date.
1/5/12:
Today, I will be uploading as much data into the charts as possible during the class period. Everything else will be finished after school today and this weekend. Unfortunately, I believe that I still have too much to finish to be able to turn in my finished product tomorrow for the extra credit, but the weekend will give me plenty of time to finish it up and turn it in on Monday at the beginning of class.
During the period today, I accidentally put the information in the wrong graphs, so I will have to fix it at home today.
1/5/12:
Fixing the incorrect data tables, calculating percents, doing some organization and beginning to write all of my discussions and conclusions. Also, I calculated my success rates and compared them. Facial Idiosyncracies proved to be the best.
1/7/12:
Today, I will be finishinng my project! I will be typing all of my final conclusions, dicussions, and analysises, orgainizing, and uploading my figures (graphs and pictures).
I am dealing with the formatting the best I can at this point. Everytime I save, it goes back to the way it does before I last saved.
Data TablesT= TrueL= Lie
Table 1: Control Statements from all 10 participants
(BPM)
Speech
*Touches her face
*Makes noises to hide laughter
*Looks around the room intently
discomfort
*Raises eyebrows
*Leans forward
and "um...")
*Mumbles/ slurs words
*fidgets with hands
*avoids eye contact
consumed a Mountain Dew soda within an hour of testing.
*Smiles and laughs when lying.
monotone
voice when lying.
Name
Arnold-Scerbo
Vebares
Iaria
Lyons
Haynie
Gildea
Edwards
Zeplo
Nelson
Lyons
Name
Arnold-Scerbo
Vebares
Iaria
Lyons
Haynie
Gildea
Edwards
Zeplo
Nelson
Lyons
Table 4: Round 2 Questioning- Speech
Name
Arnold-Scerbo
Vebares
Iaria
Lyons
Haynie
Gildea
Edwards
Zeplo
Nelson
Lyons
Name
Arnold-Scerbo
Vebares
Iaria
Lyons
Haynie
Gildea
Edwards
Zeplo
Nelson
Lyons
Table 6: Round 3 Questioning- Pulse
Name
Arnold-Scerbo
Vebares
Iaria
Lyons
Haynie
Gildea
Edwards
Zeplo
Nelson
Lyons
Name
Arnold-Scerbo
Vebares
Iaria
Lyons
Haynie
Gildea
Edwards
Zeplo
Nelson
Lyons
Success Rates: (# of lies correctly detected/ Total # of lies)x 100 = % of success
Round 1 (Facial Idiosyncracies)- Success Rate: 25/30= 0.8333 x 100= 83.33%Round 2 (Speech)- Success Rate: 20/30= 0.6667 x 100= 66.67%
Round 3 (Pulse)- Success Rate: 12/30= 0.4 x 100= 40%
The most successful test was Facical Idiosyncracies, which proved my hypothesis wrong.
AnalysisExperimental Background:
To lie is to present a statement that is not the complete and whole truth. The average person lies several times a day about a wide variety of things. The spectrum of lying ranges from little white lies to deep, malicious lies intended to hurt someone. Testing a lie can be done in several ways. These may include: most popularly, a polygraph test, analysis of body language or facial idiosyncracies, or the way the speak and explain themselves. None of these are a definite way to spot lying. A professional grade polygraph test has a 40% margin of error. Neither of the other trials even have a set rate of error or success because they are so difficult to test due to the fact that they are situational.
My project tested the accuracies of several lie detection techniques. I tested these less-used techniques against a make-shift polygraph test (blood pressure/ pulse monitor). I rounded up 10 test subjects between the ages of 13-19 years old and tested their lies against 3 variables: Facial idiosyncracies and body language, speech, and pulse. Each variable was tested with a set of 10 questions. The test subjects were prompted to lie for 3 questions per set of 10. I then had to use each of these techniques to determine which answers were lies upon my questioning them. According to my research, the pulse test should have been the most accurate, while in fact, it was the least.
Discussion:
This experiment was designed to compare three lie detectors and determine which was most successful. By testing the techniques of facial/body language, speech, and pulse, I determined that facial and body language are the most successful way to determine lying. You can see this at the end summarizations for Tables 3, 5, and 7 and Figure 1. This, although, goes against all of my research which stated that pulse and blood pressure would be most accurate. In fact, the pulse test proved to be the least effective by far, which can be seen in the summarization of Table 7 and Figure 1.
My reasoning for being most accurate for facial and body language are that I knew my participants very well. I am accustomed to their idiosyncracies and their lying patterns. My results may have been drastically different if I had been testing people that I have never met. At that point, I would have to completely rely on my very basic knowledge of lying and what it does to face and body language. Instead, I could use prior knowledge of my participants to aid me in my decision making process. Because of this, my data may have been a little clouded. My data may have also been roughed up because of the quality of my equipment. My pulse monitor way by no means professional-grade (Figure 2). It often took a few tries to get a reading on a participant's pulse. Because of this, it was also easily manipulated. After my participants found this to be true, they silently began taking advantage of it. As you can see when comparing Tables 6 and 7, I sometimes did not detect any lies from participants. It was very hit or miss. If I were to redo my experiment, I would definitely invest in a more proper pulse monitor.
Another possible margin of error could be that I let my test subjects review and hold the question sheets before and while I was questioning them (Figure 3). This way, they had time to decide which they were going to lie for and how. They did not have to lie on the spot, which made it easier to manipulate their features, body language, and breathing (which effected their pulse). This may have effected my data in more of a way than I had originally predicted. I would also change this about my project to improve it. The added stress of having to come up with a lie on the spot may have an affect on the subject's pulse, features, and body language. Depending on how much of an impact this had on each person, it could change my end results completely.
The final thing I would change to improve the project is that I would test more participants or expand my age group. Obviously, by testing more people, I could achieve a more accurate average for each technique. But for the purposes of this project, I believed 10 would be enough. If I were to continue this project next year, I would like to see if testing a bigger demograpphic would vary my end results (which can be seen in Figure 1) Originally, I chose the age group that I did for reasons that applied to the way I was originally going to go about the project. But as I changed my ideas it became irrelavent to my question. Maybe it is more difficult to trigger a spike in blood pressure in young people as opposed to an older age group. They were all relatively close together (Table 1) and did not vary much through my testing. That could be another aspect reasoning for not being able to detect a majority of the lies in the pulse section and its surprisingly low success rate of 40% (Seen in Tables 6 and 7 and Figure 1).
I predicted that speech and face/ body language would be close to each other. But as I documented my control statements (Table 1) and realized that few of my participants had definite speech ticks. It was here that I began to think that speech would fall just behind facial ticks, which it did. But, if I did not know my participants so well, they could possibly have lied about their control statements. That could be a potential margin of error for testing stangers, and I would definitely make that a part of my consent/ agreement form if I were to take that step with this experiment later.
To conclude, by testing the techniques of facial ticks and body language, speech ticks, and pulse spikes on 10 participants, I found that the best way to detect lying is through facial ticks and body language (as seen in Figure 1). There were several flaws in my experiment that I would change if I were to continue, and I would also invest in more professional equipment. Now that I already have a base in the experiment, my predictions for its continuation stand with my final results here. These can be seen in Figure 1 and the math is shown in the summarization after Table 7.
Application:
The results in the experiment could be very beneficial. That ranges from a Detective or police officer, a lawyer, to someone who is just looking to spot a lie in a story. By knowing how to effectively tell if a person is lying to you, it is easiest to be able to find the ticks that confirm or deny their statement. Even more, if you personally are trying to lie your way out of a situation, you would know what to avoid as to not be caught, such as someone in the who has gone undercover. The study could go further to other ways of detection which could be compared to the ones tested here. That way, there would be even more of a variety of accuracy. All in all, this experiment can pertain to anyone who ever has needed to spot a lie or has ever needed to lie their way out of a situation.
Conclusion:
My hypothesis stated that the most effective way to spot a lie was that their pulse would spike. As you can see through all of my data tables and Figure 1, this hypothesis is incorrect. The most accurate way to spot a lie, as determined by this experiment, it a person's facial ticks and body language. After testing 10 participants on all three of the techniques, facial ticks had the highest success rate by far. The percent of success being 83.33%, 66.67% for speech, and 40% for pulse. Despite the flaws in this experiment, it is safe to conclude that by analyzing a person's facial features and body language, it will most effectively detect if a person is lying or not.
Figures
Figure 1
I'm not sure why it wont let me just have the picture of the graph in here but I suppose that it is because this is a Pdf file. It opens to the graph when you double click on it. It is my success rate comparison graph.
Figure 2
Figure 3
Sources:
Ps: This is not the only place where it refuses to save the formatting. It changes font size, color, and placement all over the document.